Growth heterogeneity is a recurrent problem in fish rearing systems. It is the result of complex
interactions between numerous factors: feeding methods (Gelineau et al.,1998), social interactions
(Jobling, 1995, Greaves and Tuene, 2001), environmental condition (Jobling and Baardvick, 1994)
and genetics (Tiira et al. 2006). In order to shed new light on this subject, we developed an
individualized based model (named Mo.B.I.Fish) to reproduce in silico the biological phenomena. This
model was subdivided into two sub-models. The first one is a bio-energetic model that allow simulating
individual growth from individual food intake. The second sub-model is used to simulate social
interactions among fish. The interactions were simulated through a multi-agents system that
reproduced fights between fishes. The factors of influence in these contests were selected from
ethological studies. The simulated result of the fight was a function of weight, experience of each fish,
genetic and random effect. The result of the fight had direct effect on individual intake (in order to
simulate the dominance). The output of this model was the individual growth trajectory, from which
were extracted the mean body weight of the population, the coefficient of variation (CV) and the
Spearman rank correlation coefficient. Experimental data were used to calibrate and validate the
model (based on the two previous outputs): the first evaluation was conducted on sea bass,
Dicentrarchus labrax, reared with different initial heterogeneity. The second one was conducted on
Eurasian perch, Perca fluviatilis, which were reared under a factorial design of 3 factors (initial
heterogeneity, feeding amount and feeding distribution) with 2 modalities (8 experimental units). After
calibration and validation of this model, we were able to estimate the parameters and test the
likelihood of our hypotheses. The genetic variations were not used. Size had little or no effect on the
outcome of the fight (unless we added an extra-hypothesis that stipulated that the fights only occurred
between fish of similar size), however a simulation in which the outcome of the fight was mainly ruled
by the experience of each fish (also called winner effect and looser effect) led to very realistic results.
The growth of the CV and its stagnation, which is an emergent property of this model, is accurately
simulated. In the future, more comparison between the model and experimental data could validate
these first results and allow extending the model to other species, such as chicken and pigs.